Introduction
Colorectal cancer (CRC) is the third most incident cancer (10 %), and the second cause
of cancer-related deaths (9.4 %) worldwide [1 ]. Among detected CRC cases, 6 % to 7 % occur after colonoscopy as interval CRC representing
a failure in the screening and surveillance program [2 ]
[3 ]
[4 ]. Missed lesions and incomplete resection of colorectal polyps are the main risk
factors for the development of interval CRC [5 ]
[6 ]
[7 ]
[8 ]
[9 ]. Ensuring complete endoscopic polyp resection is therefore important to reduce incidence
of CRC and its related morbidity and mortality [6 ]
[10 ]
[11 ]
[12 ].
Recent guidelines from the US Multi Society Task Force (USMSTF) and European Society
of Gastrointestinal Endoscopy (ESGE) suggest use of cold snare polypectomy (CSP) for
1- to 9-mm polyps [8 ]
[13 ]. CSP with or without SI was suggested as a potential primary approach for 10- to
19-mm polyps in USMSTF guidelines and as an alternate approach in cases where there
is high risk of deep mural injury in ESGE guidelines [8 ]
[13 ]. A recent meta-analysis indicated that incomplete resection rates (IRRs) when using
CSP is high but comparable to IRRs when using hot snare polypectomy (HSP) [14 ]. A previous study conducted by our research group showed that CSP without SI for
4- to 20-mm colorectal polyps resulted in 19 % IRR [15 ]. In contrast, another study showed that wide-field cold polypectomy with SI was
associated with a much lower IRR (1.2 %) when used for resection of large (> 10 mm)
serrated lesions (SLs) [16 ]. We therefore hypothesized that a technique combining wide-field CSP with routine
submucosal injection (CSP-SI) might result in low IRRs for 4- to 20-mm non-pedunculated
colorectal polyps.
Patients and methods
This study has been reported according to the guidelines for reporting non-randomized
pilot and feasibility studies [17 ].
Study design and patients
We conducted a prospective multi-endoscopist single-center cohort study. Patients
45 to 80 years undergoing elective screening, surveillance, or diagnostic colonoscopy
at the Montreal University Hospital Center (CHUM) from January to October 2021 were
enrolled. Exclusion criteria were: known inflammatory bowel disease; hereditary CRC
syndromes; poor general health (defined as American Society of Anesthesiologists [ASA]
classification > 3); coagulopathy (defined as international normalized ratio ≥ 1.5
or platelets < 50); poor bowel preparation (defined as Boston Bowel Preparation Score
< 6, or < 2 in any colonic segment); emergency colonoscopies or hospitalized patients
and pregnancy. The study was approved by the Montreal University Research Center Institutional
Review Board (CER 20.111) and registered at ClinicalTrials.gov (NCT04548947). All
patients signed a written informed consent form for study participation.
Colonoscopy and polypectomy procedures
All patients underwent bowel cleansing using a standard regimen (split-dose polyethylene
glycol) before the colonoscopy. Patient antithrombotic and anticoagulation therapy
was managed according to the American Society of Gastrointestinal Endoscopy (ASGE)
guidelines [18 ]. The quality of bowel cleansing was assessed by the endoscopist during the procedure
according to the Boston Bowel Preparation Scale [19 ]. Colonoscopies were performed by five board-certified gastroenterologists using
high-definition colonoscopes (Olympus 190 series; Olympus, Center Valley, Pennsylvania,
United States of America). The size of all detected polyps was estimated using a catheter
tip of a closed snare (approximately 2.5 mm). Polyps were characterized for morphology
according to the Paris classification [20 ] and location within the colon (proximal colon from cecum to splenic flexure, and
distal colon from descending colon to rectum).
All detected 4- to 20-mm non-pedunculated polyps were resected using submucosal injection
(ORISE; Boston Scientific, Marlborough, Massachusetts, United States) followed by
wide-field cold snare resection (Captivator cold, 10 mm; Boston Scientific) ([Video 1 ]). Polyps that could not be removed using CSP-SI (Cold snare polypectomy with submucosal
injection) were removed using a hot snare or another standard resection technique
according to the endoscopist judgment. After polypectomy, all resection margins were
visually inspected using white-light endoscopy and/or narrow-band imaging for detection
of any polyp remnants. Any remaining tissue was removed using a cold snare until polyp
removal was visibly complete and polypectomy field was extended by at least 3 mm to
achieve wide-field resection. After complete resection of all visible polyp tissue,
biopsies were taken according to original polyp size: two margin biopsies from opposite
sides of the resection defect for 4- to 9-mm polyps, and four margin biopsies from
the four quadrants for 10- to 20-mm polyps.
Video 1 Injection and cold snare polypectomy technique.
Post-colonoscopy follow-up procedure
Post-procedure complications were assessed in the endoscopy suite, and the recovery
room until patient discharge. All patients were contacted by telephone 14 days after
the colonoscopy to assess for any adverse events (AEs) (delayed bleeding; abdominal
pain; perforation).
Histopathology evaluation
Board-certified pathologists assessed polyps for histopathology (according to the
2019 World Health Organization guideline) [21 ], and classified them as hyperplastic polyps, adenomas (e. g., villous adenomas,
tubulovillous adenomas, tubular adenomas), SLs, high-grade dysplasia or cancer. Margin
biopsies were assessed separately to determine the presence or absence of any remnant
polyp tissue, and thus confirm complete or incomplete resection.
Study aims
The primary aim was the IRR of 4- to 20-mm non-pedunculated colorectal polyps following
CSP-SI. Incomplete resection was defined as any margin biopsy containing tissue remnant
corresponding to the pathology of the resected polyp (e. g., adenoma remnants in the
margin biopsies after resection of an adenoma). Other outcomes included IRR stratified
by polyp size, histology, location, endoscopist experience; technical success rates
(defined as visually complete resection using CSP-SI exclusively without conversion
to HSP); Intraprocedure bleeding rate (IPB); complication rate including delayed bleeding,
abdominal pain, perforation; en bloc resection rate; visual incomplete resection after
the first polypectomy cut.
The technical success rate was calculated as the number of polyps successfully resected
with CSP-SI divided by the total number of polyps resected, excluding protocol deviations
involving initial polyp removal by a technique other than CSP-SI. The conversion rate
was defined as the proportion of polyps resected using a method other than CSP-SI
after initial attempt with CSP-SI. Easy resection was defined as complete dissection
of the submucosal plane upon closing the snare for the first time. Difficult resection
was defined as not being able to dissect and complete polypectomy upon initial closing
of the snare or physical manipulation such as “guillotine” technique being required.
IPB was defined as any bleeding that occurred during endoscopy at the polypectomy
site. IPB was defined as mild if no endoscopic treatment was required, and significant
if endoscopic hemostasis was required. Delayed bleeding was defined as any bleeding
post-procedure that necessitated a hospitalization or emergency room visit. Endoscopists
with a practice that includes doing routinely referred endoscopic mucosal resection
(EMR) cases prior to the study participation were considered experienced and those
with no routine EMR practice were considered not experienced.
All polyps that were not adenomas, hyperplastic polyps, or SLs (e. g., those that
were hamartomas, lymphoid aggregates, inflammatory polyps, subepithelial lesion, mucosal
prolapse) were not included in the IRR analysis. Adenoma detection rates, calculated
as the proportion of examinations with adenomas on the total number of examinations,
and IRR were calculated for each participating endoscopist.
Sample size and statistical analysis
A previous study conducted by our research group showed that CSP without SI for 4-
to 20-mm colorectal polyps resulted in high IRR (24/128 = 19 %) (12.4 %-26.6 %) [15 ]. We considered a reduction to 8.5 % to these previous findings as clinically significant
for demonstrating clinical superiority of CSP-SI. A one-proportion chi-squared test
with a 0.05 two-sided significance level (80 % power) allows the detection of such
a difference or more when polyp sample size is 168. We expected a polyp detection
rate of 40 %, therefore a sample size of at least 420 patients was deemed necessary
for the study. To adjust for any study dropouts (e. g., inadequate bowel preparation,
withdrawal of consent), we increased the overall sample size and enrolled a total
of 435 patients. The study was not powered to detect variation in IRRs between the
five endoscopists.
IRR was calculated as the number of polyps with positive margins determined by histopathological
examination divided by the total number of polyps and is presented with 95 % confidence
intervals (CIs). The IRR was calculated the 95 % CI considering the varying number
of study polyps per patient, with a Bootstrap of 10 000 samples, stratified by patient
record ID. Descriptive statistics are presented as numbers and frequencies for categorical
variables, and as mean (standard deviation [SD]) or median (range) for continuous
variables with normal and non-normal distribution, respectively.
To assess the predictors of incomplete resection, we used generalized estimating equations
(GEEs) with logit link and exchangeable correlation matrix. The GEEs account for clustering
of polyps within patients (as a patient may have more than one polyp). To construct
the GEEs, we first conducted univariable logistic regression models and variables
significant in these models at the P < 0.05 were entered in the GEEs. The potential predictors considered included: polyp
characteristics (i. e., size 4 to 5 mm vs. 6 to 9 mm vs. 10 to 20 mm), surface morphology
(flat vs. non-flat), location (proximal vs. distal), histopathology (hyperplastic;
SLs vs. adenomas (reference category); polypectomy characteristics such as polyp resection
type (en bloc vs. piecemeal), ease of resection (easy vs. difficult); endoscopist-related
characteristics (i. e. level of experience with CSP-SI (experienced with EMR vs. inexperienced
with EMR), case volume, cecal intubation rate, and adenoma/polyp detection rate).
Results were reported as odd ratios (ORs) with 95 % CIs. P < 0.05 was considered statistically significant. Additional analyses were also performed
to analyze the association between endoscopists (performing ≥ 14 polypectomies for
4- to 20-mm polyps) and incomplete resection, considering the endoscopist with the
lowest IRR as reference. All statistical analyses were performed using SPSS statistics
version 27 (IBM Corp., Armonk, New York, United States).
Results
Patient, procedure, and polyp characteristics
A total of 429 patients (median age 65 years; 47.1 % female) were enrolled in the
study. After applying inclusion and exclusion criteria, a total of 429 patients were
enrolled in the study, of whom 128 patients with non-pedunculated polyps sized 4 to
20 mm were included in the final analyses ([Fig. 1 ]). The included patients had 206 potentially eligible polyps and of these, 183 met
the inclusion criteria and were included in the final analyses ([Fig. 1 ]). Of the included polyps, 87 (47.5 %) were 4 to 5 mm, 64 (35.0 %) were 6 to 9 mm
and 32 (17.5 %) were 10 to 20 mm. The majority of the polyps, 129 (70.5 %) were adenomas,
25 (13.6 %) SLs and 29 (15.8 %) HPs. Detailed patient, procedure, and polyp characteristics
are provided in [Table 1 ].
Fig. 1 Patient enrollment flowchart.
Table 1
Patient, polyp and procedure characteristics.
Patients, n
429
Age, median (IQR) [range], years
65 (12) [45–80]
Sex, male, n (%)
227 (52.9)
ASA, n (%)
129 (30.1)
277 (64.6)
23 (5.4)
Anticoagulant and antiplatelet therapy, n (%)
93 (21.7)
61 (14.2)
32 (7.5)
Adenoma detection rate, n (%)
172 (40.1)
Adequate bowel preparation, n (%)
386 (90.0)
Polyps, n
183
Mean size [SD]
6.82 [3.5]
Location, n (%)
121 (66.1)
62 (33.9)
Paris classification
139 (76.0)
41 (22.4)
3 (1.6)
Size
87 (47.5)
64 (35.0)
32 (17.5)
Histology
129 (70.5)
120 (65.7)
1 (0.5)
6 (3.3)
2 (1.0)
25 (13.7)
29 (15.8)
ASA, American Society of Anesthesiologists; IQR, interquartile
range; HGD, high-grade dysplasia.
Technical outcomes and Incomplete resection rates
CSP-SI was reported to be technically easy in 92.6 % polypectomies (186/204). The
rate of conversion to HSP was 2.5 % (5/204). Details of polypectomy procedure outcomes
are shown in [Table 2 ].
Table 2
Technical outcomes of CSP-SI.
All polyps
4–5 mm
6–9 mm
10–20 mm
P value
Type of resection[1 ], n (%)
98 (48)
60 (61.2)
34 (34.7)
4 (4.1)
< 0.001
104 (50.9)
37 (35.6)
35 (33.7)
32 (30.7)
Ease of resection[2 ], n (%)
186 (91.2)
90 (48.4)
65 (34.9)
31 (16.7)
0.006
11 (5.4)
6 (54.5)
4 (36.4)
1 (9.1)
Visible residual polyp after first cut
124 (60.8)
75 (0.8)
39 (31.5)
10 (8.1)
< 0.001
80 (39.2)
22 (27.5)
32 (40)
26 (32.5)
CSP-SI, cold snare resection with submucosal injection; HSP, hot snare polypectomy.
1 Unknown resection technique for two polyps.
2 Five HSP conversions and two non-resected polyps (N/A).
Of the 183 polyps included in the final analysis, seven were incompletely resected,
with an overall IRR of 3.8 % (95 % CI 2.7–5.5 %). The GEE analyses revealed that the
IRR was statistically significantly higher for SLs (16.0 %) (OR 12.11 [95 %CI 2.1–71.20]);
and hyperplastic polyps (3.4 %) (OR 2.27 [0.19–28.02]) compared to adenomas (1.6 %)
(P = 0.016). IRR for polyps 4 to 5 mm was 2.3 % vs 6.3 % for polyps 5 to 9 mm and 3.1 %
for polyps 10 to 20 mm. IRR for polyps < 10 mm was 4.0 % (95 %CI 2.6–6.0). There was
no statistically significant difference in IRR between the polyp size groups (P = 0.467) ([Table 3 ]). IRR was higher for en bloc resection when compared to piecemeal resection although
not statistically significant (5.8 vs 2.1 %; OR 2.9 [95 %CI 0.6–15.5]; P = 0.203). Difficulty in use of CSP-SI had statistically significantly higher IRRs
compared with ease of use (18.2 vs 2.9 %) (OR 7.47 [95 %CI 1.28–43.75]; P = 0.026). Endoscopists inexperienced with EMR had statistically significantly higher
IRRs compared to those who were experienced (10.5 vs 2.1 %; OR 5.75 [95 % CI 1.30–25.40];
P = 0.021). Endoscopists included in the study had varying results for endoscopic quality
metrics, however, these did not seem to be associated with IRRs ([Table 4 ]).
Table 3
Incomplete resection rates after CSP-SI of colorectal polyps and associated predictors
IRR, n (%)
Univariable OR [95 % CI]
P value
All (n = 183)
7 (3.8)
Size, n (%)
2 (2.3)
1.00
0.467
4 (6.3)
2.07 [0.22–19.3]
1 (3.1)
0.72 [0.06–8.33]
Histology, n (%)
2 (1.6)
1.00
0.016
4 (16.0)
12.09 [2.08–70.24]
1 (3.4)
5.33 [0.55–51.27]
Location, n (%)
5 (4.1)
1.00
0.761
2 (3.2)
1.29 [0.24–6.87]
Type of resection, n (%)
2 (2.1)
1.00
0.203
5 (5.8)
2.9 [0.55–15.52]
Ease of resection, n (%)
5 (2.9)
1.00
0.026
2 (18.2)
7.42 [1.26–43.65]
Visible residual polyp after first cut
5 (4.7)
1.00
0.484
2 (2.6)
0.55 [0.10–2.92]
Level of endoscopist experience
3 (2.1)
1.00
0.021
4 (10.5)
5.57 [1.19–26.05]
CSP-SI, cold snare resection with submucosal injection; IRR, incomplete resection
rate; OR: odds ratio.
Table 4
Endoscopist variability according to ADR, CIR and IRR (for 4– to 20-mm polyps).
Endoscopist
ADR (%)
CIR (%)
Withdrawal time
Study polyps removed (N = 206)
Average volume of submucosal gel injected per polyp (mL), median (IQR)
IRR (N = 183)
A
46.8
95.9
9 (6,14.25)
94
4 (2,5)
0 (0)
B
27.5
90.1
9 (7,14.25)
31
3 (2,6)
3 (11.5)
C
45
82.5
13 (9,23)
48
3 (2,4)
1 (2.3)
D
35.1
93.2
9 (6.5,15)
15
3 (2,7.5)
1 (8.3)
E
43.6
90.9
10 (8,15)
18
2 (1,4)
2 (12.5)
Group 1[1 ]
45
92.8
10 (7,16)
160
3 (2,5)
3 (2.1)
Group 2[2 ]
30.7
91.5
9 (7,15)
46
3 (2,6)
4 (11)
ADR, adenoma detection rate; CIR, cecal intubation rate; IRR, incomplete resection
rate.
1 Group 1, experienced endoscopists A, C & E.
2 Group 2, non-experienced endoscopists B & D.
Adverse events
Mild self-limiting IPB occurred in 69 patients (53.9 %, 69/128) [95 % CI 44.87–62.75 %],
four of which underwent clip placement (3.1 %, 4/128) [95 % CI 0.86–7.81 %]. During
the 14-day post-intervention period, clinically significant delayed bleeding occurred
in 0.8 % of patients (1/128) (95 % 0.02–4.28), abdominal pain in 17.2 % (22/128) (95 %
CI 11.09–24.85), and no perforation occurred. Out of these events, clinically significant
delayed bleeding related to CSP-SI occurred in no patients (0 %) and abdominal pain
occurred in 7 % of patients (9/128) (95 % CI 3.27–12.93 %). No serious AEs were reported
following polypectomy using the CSP-SI technique ([Table 5 ]).
Table 5
Adverse outcomes related to CSP-SI.
Patients
Confidence interval
Intraprocedure bleeding[1 ], n (%)
None
46 (35.9)
27.65–44.89 %
Mild
69 (53.9)
44.87–62.75 %
Treatment needed
4 (3.1)
0.86–7.81 %
Delayed bleeding, n (%)
0 (0)
0.00–2.84 %
Abdominal pain, n (%)
9 (7)
3.27–12.93 %
Perforation, n (%)
0 (0)
0.00–2.84 %
CSP-SI, cold snare resection with submucosal injection; CI, confidence interval.
1 Bleeding occurring during the procedure. Unknown bleeding for seven patients and
not applicable to two patients; binomial Clopper-Pearson confidence interval.
Discussion
In this prospective exploratory cohort study, we found that using CSP-SI was associated
with low IRR (3.8 %). CSP has seen recent increased popularity due to its favorable
safety profile, with current American and European guidelines in many cases recommending
its use over HSP or cold forceps [8 ]
[13 ]. There have been many studies evaluating IRRs for standard CSP with varying reported
rates of 7 % to 30 % for varying polyp sizes [22 ]
[23 ]
[24 ]
[25 ]
[26 ]
[27 ]. A recent meta-analysis of IRRs found 17.3 % [95 % CI 14.3–20.3] IRR for CSP of
1- to 20-mm polyps which was similar to IRRs for HSP [14 ]. IRR therefore remains very high in the literature and could contribute to interval
CRC cases, with studies attributing 20 % to 30 % of interval CRC to incomplete resection.
Interval CRC is typically found in colon segments of previous polypectomy sites [28 ]
[29 ]
[30 ]
[31 ]
[32 ]
[33 ]. It is, therefore, important to reduce IRRs to ensure proper efficacy of CRC screening
programs. In our study, SI and wide-field resection was used to counteract the high
IRRs observed for standard CSP with significantly lower incomplete resection than
the reported literature on standard CSP.
The use of EMR-style techniques for CSP of 4- to 20-mm polyps is very sparce in the
literature with varying results. Two studies using normal saline SI found 5.9 % IRR
when using cold EMR (C-EMR) in 6- to 20-mm polyps, and 7.2 % in 6- to 10-mm polyps
[34 ]
[35 ]. Another in 3- to 10-mm polyps found that SI did not improve the resection depth
[36 ]. One study in 10- to 14-mm polyps using glycerol SI found excessively high (36.2 %)
IRRs [37 ]. Gel injection solutions could offer more long-lasting dome-shaped cushion elevation
underneath the mucosa compared to saline which is crucial for polypectomy [38 ]. In a previous randomized controlled trial (RCT), it was found that the use of a
viscous gel injection reduced the resection times as well as the number of pieces
significantly compared to NS [39 ]. Saline and HES are used as an off-label polypectomy aid, and they are time-consuming
to prepare when MB or other agents are added in the solution. In our practice, we
found pre-dyed gel injections quicker and easier to use compared to saline or HES. Although
we did not measure injection times in all polypectomies, injection time was generally
less than 1 minute and did not significantly impact procedure times rendering this
option viable even for smaller polyps. Given the increased costs associated with utilization
of gel-based SIs, head-to-head comparison to cheaper alternatives such as NS are required
before widespread adoption.
A recent RCT compared cold snare, C-EMR, hot snare or hot EMR for resection of 6-
to 15-mm non-pedunculated polyps. The study found that cold snare had the lowest IRR.
However, the study was designed to demonstrate non-inferiority between the four different
polypectomy techniques and thus included only 286 polypectomies randomized into the
four groups and only seven incomplete resections were found across all four groups
[40 ]. Furthermore, all polypectomies were performed by expert endoscopist with extensive
EMR experience and polypectomy techniques were not standardized (i. e. different SI
agents and different snare types were used within the same group). In contrast our
study was comprised of a group of EMR experts and non-experts all using a standardized
polypectomy technique. We found in our study that EMR experts tended to have lower
IRR compared to non-experts. The use of wide-field EMR where the resection margins
are systematically expanded might also have contributed to the results seen in our
study when comparing the IRR found (3.8 %) compared to a much higher IRR found (19 %)
in a previous study conducted by the same group of endoscopists using standard CSP
alone and in comparison to hot snare study with similar methodology ([Table 6 ]). However, further RCTs are required to demonstrate superiority regarding IRR between
CSP, CSP-SI and hot snare-based techniques. It is possible that the wide-field resection
aspects when using CSP would have also improved IRR. Indeed, recent publications have
highlighted that wide-field resection allows for a better completion of the resection
of lesions measuring less than 10 mm, and a lower IRR [41 ]. Furthermore, using the different technical approaches (e. g., SI versus no injection)
needs to be evaluated for different polyp entities (i. e., SLs).
Table 6
Historical cold snare and hot snare polypectomy outcomes with or without submucosal
injection.
Cold snare with submucosal injection
Cold snare without submucosal injection [15 ]
1 ]
Hot snare without submucosal injection [42 ]
2 ]
Polypectomy achieved as per protocol, n (%)
199 (97.5)
128 (70.3)
346 (100)
Ease of resection[3 ], n (%)
186 (90.3)
136 (77.7)
222 (64.2)
N/A
N/A
75 (21.7)
11 (5.3)
22 (12.6)
45 (13)
No (conversions)
5 (2.4)
17 (9.7)
0 (0)
Deviation
2 (1.0)
0 (0)
0 (0)
Volume of submucosal injection, median (IQR) mL
3.87 (2–5)
0 (0)
0 (0)
No. polyps included in the primary outcome analysis, n
183
128
346
Incomplete resection, n (%)
7 (3.8)
24 (19)
35 (10.1)
En bloc method, n (%)
86 (47)
128 (90.1)
286 (82.7)
Residual tissue after resection, n (%)
76 (41.5)
22 (15.4)
N/A
Bleeding[4 ], n (%)
85 (46.4)
65 (38.5)
261 (97)
85 (46.4)
96 (56.8)
N/A
4 (2.2)
8 (4.7)
8 (3)
CSP-SI, cold snare resection with submucosal injection; CSP, cold snare polypectomy;
IQR, interquartile range.
1 Missing bleeding rate for 13 patients.
2 Bleeding rate calculated on a total of patients (n = 269).
3 Resection methods for two polyps are not applicable for CSP-SI study (N/A = 2).
4 Unknown bleeding for seven patients and not applicable for two patients.
Polyp histology was indeed associated with IRR outcomes. SLs were found to have significantly
higher IRR (16.0 %) when compared with adenomas (1.6 %) or hyperplastic polyps (3.4 %)
(OR 12.11, [95 % CI 2.10–71.20]; P = 0.016). This finding can be explained by SL’s inconspicuous and usually indefinite
border structure [42 ]. There is little data on CSP specifically for 1- to 20-mm SLs. One study in 10-
to 20-mm polyps found 44.3 % IRR, however, recurrence rates were very low 4.9 %, in
contrast to another study performed by expert endoscopists using SI where only 2 %
were incompletely resected [43 ]
[44 ]. Endoscopist expertise in conjunction with EMR-style techniques could therefore
contribute toward reducing IRR. In our study, experienced endoscopists had statistically
significantly lower IRR (2.1 vs 10.5 %) (OR 5.8 [95 %CI 1.30–25.40]) (P = 0.021), with a moderate IRR for SLs when compared to the literature when expert
and non-expert endoscopists are combined (16.0 %).
Studies directly comparing CSP to HSP have shown increased IPB when using CSP; however,
CSP had very low rates of delayed bleeding, which is a more significant outcome as
it can be associated with significant morbidity compared to IPB which can be addressed
during the procedure [35 ]
[45 ]
[46 ]
[47 ]. One study directly comparing C-EMR to conventional EMR in > 20 mm polyps found
no bleeding and no perforation in the C-EMR group compared with 5.1 % delayed bleeding
and 0.6 % perforation in the EMR group [48 ]. We conducted systematic 14-day post-intervention period calls to all patients in
the cohort. Furthermore, we conducted systematic assessment of IPB documenting non-significant
self-limited bleeding. During the 14-day post-intervention period, clinically significant
delayed bleeding related to CSP did not occur, highlighting the safety profile of
this technique. Of the patients that experienced abdominal pain, two patients had
non-study polyps cut with HSP.
Studies directly comparing CSP to HSP have shown increased IPB when using CSP; however,
CSP had very low rates of delayed bleeding which is a more significant outcome as
it can be associated with significant morbidity compared to IPB which can be addressed
during the procedure [35 ]
[46 ]
[49 ]
[50 ]
[51 ]. There were no CSP-SI-related serious AEs (i. e., no perforations) in our study
and no clinically significant delayed bleeding. CSP-SI is, therefore, very safe for
4- to 20-mm polyps. Some publications have reported histological changes caused by
the gel injection utilized in our study, although our pathologists have not found
any immediate changes in our large cohort pathology samples. The previously described
findings were found in a cohort of less than 60 patients [52 ]
[53 ]
[54 ]. Therefore, a larger analysis would need to be conducted in comparison to another
injection to associate these changes with the SI and follow-up will be planned in
our cohort to assess these changes. Although a specific gel solution was used in our
study, this does not preclude utilizing wide-field resection with SI with other solutions.
Further studies are required to demonstrate generalizability using a wide range of
injection solutions.
To our knowledge, very few studies reported on CSP-SI in 4- to 20-mm and no study
reported on the use of gel submucosal injection substance. The granular reporting
of polyp sizes and histology is a strength of our study. The inclusion of experienced
and non-experienced endoscopist is an added strength and improves its generalizability
to routine endoscopic practice. Our study, however, has several limitations such as
the non-experienced endoscopists that contributed relatively fewer polypectomies to
the study. The single-centered nature of the study could limit its generalizability
and also larger RCTs are required. Any commercial injection solution such as the SI
gel is certainly more expensive than a saline solution or HES although the latter
are not US Food and Drug Administration-labeled for that use. Although we were able
to observe a low IRR of 3.8 % with this injection, to confirm its beneficial use it
would be important to pursue a comparative RCT study with saline in which the IRRs
of non-pedunculated polyps sessile SLs are evaluated. Because our endoscopists resected
47 % of the lesions piecemeal, another limitation is the risk of local recurrence
rate post-piecemeal resection. For the follow-up study, we plan a 2-year follow-up
of this cohort to report if there were any recurrences for the seven IRRs (3.8 %)
reported in this study. During that follow-up, we will also observe and report any
potential appearance of histological changes that were not observed in the present
study.
Conclusions
In conclusion, CSP-SI results in very low (3.8 %) IRRs for 4- to 20-mm polyps. CSP-SI
could be considered a safe and effective approach to remove 4- to 20-mm colorectal
polyps; however, comparative studies of CSP without SI are required to confirm these
results.
